Dyna-P: Placement-aware Dynamic Partitioning for Lightweight Applications with Modern GPUs

Read the full article See related articles

Listed in

This article is not in any list yet, why not save it to one of your lists.
Log in to save this article

Abstract

Efficient GPU resource sharing is critical in dynamic cloud-based environments, particularly for lightweight HPC applications and Small Language Models, which demand partial GPU resources for execution. However, traditional scheduling frameworks fail to address intra-GPU and inter-node resource fragmentation and dynamic placement challenges arising from the heterogeneity in each application's resource demand and job completion times. This leads to resource under-utilization and scheduling delays in GPU clusters. This paper introduces Dyna-P, a novel scheduling framework designed to dynamically adjust GPU partitions to minimize resource fragmentation while improving system throughput and Makespan. Dyna-P proposes a Reconfiguration Last Policy which recognizes that workloads consisting of lightweight applications can benefit more from uninterrupted execution. Experimental results demonstrate that Dyna-P improves average throughput by up to 14.7% and reduces Makespan by 39% compared to state-of-the-art methods. These findings underscore Dyna-P’s potential to improve resource allocation rates in multi-tenant GPU environments

Article activity feed